# Dealing with large gaps in asteroseismic time series

**Authors:** Timothy R. Bedding, Hans Kjeldsen  
**arXiv:** [2303.15584](https://arxiv.org/abs/2303.15584)  
**Date:** 2023-03-27  
**Summary page:** https://www.summarizepaper.com/en/paper/2303.15584v1/

## Key points

- Asteroseismology involves studying the oscillations of stars by analyzing their power spectra.
- Long data sets are now available from space missions such as BRITE, CoRoT, Kepler, K2 and TESS.
- It is sometimes necessary to deal with time series that have large gaps, which is becoming particularly relevant for TESS.
- Solar-like oscillators have finite mode lifetimes, making it tempting to close large gaps by shifting time stamps.
- However, using actual data from the Kepler Mission, Timothy R. Bedding and Hans Kjeldsen show that this results in artificial structures in the power spectrum that compromise measurements of mode frequencies and linewidths.
- Even if modes are completely incoherent between two segments, an arbitrary shift will introduce artificial structures in Fourier space that will compromise profile fitting.
- The authors analyzed four of Zebedee's strongest modes and found that even though they were shifted arbitrarily, the power spectrum showed artificial structures that compromised measurements of mode frequencies and linewidths.
- It is important to avoid this practice when dealing with long data sets available for asteroseismology from space missions such as TESS.

## Summary

Asteroseismology involves studying the oscillations of stars by analyzing the power spectra of their flux or radial-velocity variations. With long data sets now available from space missions such as BRITE, CoRoT, Kepler, K2 and TESS, it is sometimes necessary to deal with time series that have large gaps. This is becoming particularly relevant for TESS, which is revisiting many fields on the sky every two years. Solar-like oscillators have finite mode lifetimes, making it tempting to close large gaps by shifting time stamps. However, using actual data from the Kepler Mission, Timothy R. Bedding and Hans Kjeldsen show that this results in artificial structures in the power spectrum that compromise the measurements of mode frequencies and linewidths. The Fourier transform of a long time series results in a very large array. The power spectrum is calculated with a step size that must sample the frequency resolution, which is inversely proportional to the total duration of the time series (including the gap). Even if modes are completely incoherent between two segments, an arbitrary shift will introduce artificial structures in Fourier space that will compromise profile fitting. To demonstrate this issue further, Bedding and Kjeldsen used subgiant star KIC 11137075 (also known as ‘Zebedee’), whose Kepler light curve contains more than one year of short-cadence data (1-minute sampling) showing high signal-to-noise solar-like oscillations centered at about 1700 µHz (Tian et al. 2015). They defined two segments with boundaries coinciding with short breaks each month during which data were downloaded from the spacecraft. The segment lengths were 30.1 d and 26.0 d respectively while there was a gap of 412.1 d. 

The authors then analyzed four of Zebedee's strongest modes and found that even though they were shifted arbitrarily, the power spectrum showed artificial structures that compromised the measurements of mode frequencies and linewidths. In conclusion, while it may be tempting to close large gaps in asteroseismic time series by shifting time stamps, this results in artificial structures in the power spectrum that compromise measurements of mode frequencies and linewidths. It is therefore important to avoid this practice when dealing with long data sets available for asteroseismology from space missions such as TESS.